Download A Power Re-Use Technique for Improved Efficiency Robert Langridge,

IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 47, NO. 8, AUGUST 1999
1467
A Power Re-Use Technique for Improved Efficiency
of Outphasing Microwave Power Amplifiers
Robert Langridge, Member, IEEE, Todd Thornton, Peter M. Asbeck, Senior Member, IEEE,
and Lawrence E. Larson, Senior Member, IEEE
Abstract—An improved power re-use technique is introduced
for application to outphased microwave power amplifiers. The
technique allows a significant portion of the wasted out-of-phase
components of the signal to be returned to the power supply,
resulting in substantial improvements in overall power-amplifier
efficiency. A peak re-use RF-to-DC efficiency of 63% was obtained at 1.96 GHz.
Index Terms—linear power amplifiers, LINC, outphasing amplifiers.
I. INTRODUCTION
T
HE out-phased power amplifier concept dates back to
the early 1930’s as an approach for the simultaneous
realization of high-efficiency and high-linearity amplification
[1]. It has been revived recently for wireless communication
applications under the rubric of linear amplification with
nonlinear components (LINC) [2], and many recent papers
have developed the concept further [3], [4]. The LINC concept
takes an envelope-modulated bandpass waveform and resolves
it into two out-phased constant envelope signals, which are
applied to highly efficient—and highly nonlinear—power amplifiers, whose outputs are summed. The advantage of this
approach is that each amplifier can be operated in a very
efficient “switching” mode, and yet the final output can be
highly linear—a key consideration for bandwidth-efficient
wireless communications. This is shown schematically in
Fig. 1.
However, one of the major disadvantages of the approach
is the power wasted in the summing network when the
two amplifiers are operated substantially out-of-phase, which
dramatically compromises the power-added efficiency of the
overall amplifier [5]. This problem has been addressed previously through the use of a Chireix power combiner [6], and
adaptive termination of each amplifier output depending on
the phase [7]. Though these techniques represent a significant
improvement, each still suffers from limitations, including
efficiency, bandwidth, and linearity.
Manuscript received November 30, 1998. This work was supported by
the Army Research Office Multidisciplinary Research Initiative “LowPower/Low-Power Electronic Technology for Mobile Communications”
Program.
R. Langridge, P. M. Asbeck, and L. E. Larson are with the Center for
Wireless Communications, University of California at San Diego, La Jolla,
CA 92093-0407 USA.
T. Thornton is with REMEC Wireless, San Diego, CA 92123 USA, and is
also with the University of California at San Diego, La Jolla, CA 92093-0407
USA (e-mail: [email protected]).
Publisher Item Identifier S 0018-9480(99)06085-8.
Fig. 1. Outphased power amplifier illustrating power wasted in combining
network.
A new technique is proposed here for partial recovery
of the wasted power in the summing network, similar to
recently reported results of adiabatic digital VLSI circuits [8].
The potential resulting improvement in overall power-added
efficiency is significant, without an excessive loss of design
flexibility.
II. THEORY
The outphased amplifier of Fig. 1 takes a general baseband
representation of an RF signal
(1)
is the instantaneous amplitude and
is the
where
instantaneous phase and resolves it into two constant envelope
and
through what is known as a signal
vectors
component separator (SCS), such that
(2a)
(2b)
where
The hybrid combiner of Fig. 1 takes
and
and sends them to the load resistor at the termination of
the difference port as waste heat. This process preserves the
at the output of the sum port, but reduces
envelope of
the overall power-added efficiency, especially for higher order
modulations that involve significant envelope variation. In fact,
0018–9480/99$10.00  1999 IEEE
1468
IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 47, NO. 8, AUGUST 1999
The DC value of the converter current is, therefore, approximately
(7)
for
.
where
At the same time, the impedance at the carrier frequency of
the converter can be approximated by
Fig. 2. Power recycling circuit applied to out-phased power amplifiers.
the total power-added efficiency can drop to as low as 20% for
the case of 64 quadrature amplitude modulation (QAM) [5].
In this paper, we propose to convert as much of the
wasted power as possible delivered to the difference port
termination back to the power supply. This is, in some ways,
the “mirror image” of the traditional power-amplifier design
challenge—where the goal is to convert as much of the DC
power to RF as possible. A simple example of this approach
is shown schematically in Fig. 2, where an RF–DC converter
is implemented with high-speed Schottky diodes and an optimized matching network. The efficiency of the conversion
is highly dependent on the power-supply voltage, as well
as on the series resistance of the diodes and their intrinsic
cutoff frequency and built-in voltage. Furthermore, the power
conversion is a strongly nonlinear process, requiring the diodes
to switch fully in order to achieve the best efficiency.
Assuming a simple resistive model of the diode, the power
delivered back to the power supply can be approximated by
(3)
is the power-supply voltage and
is the DC
where
value of the current from the converter. At the same time,
the power available from the hybrid is given by the familiar
expression
(8)
Since this value of the impedance varies with input power
), the converter circuit can only be ideally
(through
matched at its input at one input power level. At other powers,
there will be some loss in the recovery due to impedance
mismatch effects. If the impedance transformation is given by
, then the source voltage will be given by
(9)
and the power delivered to the detector is approximately
(10)
.
where
Hence, inserting (9) into (7) and (4) allows us to determine
the overall power re-use efficiency in terms of (5), the available
input power, diode resistance, power supply, and built-in diode
voltage.
III. MEASURED RESULTS
(4)
is the source voltage of the power-amplifier
where
output combiner. Thus, we define the overall efficiency of the
power re-use circuit as
(5)
The efficiency can be calculated by first determining the
instantaneous current through the converter, which can be
approximated by
where
is the magnitude of the peak voltage of the input
is the built-in diode voltage,
signal at the diode network,
is the series “on” resistance of the Schottky diode, and
is the carrier frequency.
AND
DISCUSSION
The power re-use circuit of Fig. 2 was fabricated with
Hewlett-Packard surface mount Schottky diodes, on a 0.5-oz
32-mil Rogers 4003 board, and was fed by a hybrid power
amplifier constructed from NE8500R599 and NE6 500 278
was used to simulate
MESFET’s. A resistive load of 11
the equivalent load of a power amplifier on the power supply,
and allowed for adjustment of the power-supply voltage. An
impedance-matching network was employed at the input to the
detector to optimize performance, and the circuit was evaluated
at 1.96 GHz, a typical personal communications system (PCS)
frequency.
Figs. 3 and 4 show the measured results for re-use efficiency
and measured VSWR as a function of the input power for
three different power-supply voltages: 3, 4, and 5 V. The best
measured re-use efficiency was found to be approximately
63% at a power level that varied with supply voltage, and
the efficiency was greater than 40% over a range of a decade
in input power. This can result in a significant increase in the
overall efficiency of an outphased power amplifier, although
there are some limitations on the use of this approach. The
experimental agreement with the simplified prediction of (5),
LANGRIDGE et al.: POWER RE-USE TECHNIQUE
Fig. 3. Measured variation of re-use efficiency at 1.96 GHz as a function
of input power level for Vsup = 3, 4, and 5 V. The agreement with (5) is
shown for comparison.
Fig. 4. Comparison of calculated and measured variation of input VSWR at
1.96 GHz as a function of input power level for Vsup = 3, 4, and 5 V.
assuming 1-dB loss in the input matching network, is excellent
over a wide range of input powers and power-supply voltages.
At low-input power levels, the efficiency drops because the
diodes are unable to turn “on” and overcome the combination
of supply voltage and built-in potential of the diode. However,
this loss in efficiency does not degrade the overall system
efficiency because most of the total RF power in the circuit
is delivered to the antenna anyway at these power levels. The
efficiency drops at higher powers because of the increasing
mismatch loss of the circuit as more power is delivered
to the diodes, and the input impedance to the converter
continues to drop. This would mean that the improvement in
efficiency could be compromised for modulation schemes that
exhibit very deep variations in envelope power on a regular
basis—like QAM.
Fig. 5 shows a plot of predicted power-amplifier system
efficiency in a LINC system with and without power reuse. For illustrative purposes, the power amplifiers in the
system are assumed to have 100% efficiency. The re-use
1469
Fig. 5. Calculated power-amplifier efficiency as a function of power delivered to the antenna for 0.8-W peak outphasing power amplifier, showing the
effect of re-use efficiency. The re-use efficiency data from Fig. 3 is included.
efficiency was varied between 0.9–0.1 to show the effects of
power re-use on the overall power-amplifier efficiency; the
data from the detector circuit is also superimposed on the
plot to compare it to the various contours of constant reuse efficiency. The “measured” data in this figure refers to
measured re-use efficiency data of Fig. 3; the power-amplifier
system efficiency was then calculated based on that data. As
can be seen, the overall power-amplifier system efficiency
can be enhanced significantly by re-using the “wasted” power
from the difference port of the 180 hybrid. The amount
of improvement will depend on the modulation scheme and
data-filtering technique applied.
The variation in measured VSWR as a function of input
power follows a similar pattern: at low input powers, the
diodes appear to be “open-circuited” and the VSWR rises
to a peak of approximately 1.8:1 and, at high input powers,
the VSWR also rises, up to a maximum of nearly 2:1. The
high VSWR at high input powers could create a limitation
on the cancellation of the signals that appear at the antenna,
creating unwanted distortion. However, the antenna hybrid
itself provides an additional 3 dB of isolation, and a circulator
could be employed to further improve the VSWR if necessary.
The bandwidth of this technique is sufficiently broad to
accommodate a variety of modulation schemes without significant degradation in the power returned to the power supply.
Fig. 6 plots the variation in efficiency using a 50% amplitude
modulation of the input signal at a 10-kHz rate. In this case,
there is little difference between this result and the continuouswave (CW) result of Fig. 3. Even wider bandwidths are
required for a CDMA case, where the typical modulation
bandwidth is increased to 1.25 MHz. In this case, the measured
results of Fig. 6 show a modest decrease of efficiency from a
peak of over 60% to just over 50%.
It should be noted that this particular power re-use circuit is
only one of several ways to approach this problem depending
on the center frequency and the bandwidth of the signal. An
alternative implementation of the power re-use circuit at lower
frequencies could be a transformer and diode pair, followed
1470
IEEE TRANSACTIONS ON MICROWAVE THEORY AND TECHNIQUES, VOL. 47, NO. 8, AUGUST 1999
[2] D. C. Cox, “Linear amplification with nonlinear components,” IEEE
Trans. Commun., vol. COM-23, pp. 1942–1945, Dec. 1974.
[3] L. Sundstrom, “The effect of quantization in a digital signal component
separator for LINC transmitters,” IEEE Trans. Veh. Technol., vol. 45,
pp. 346–352, May 1996.
[4] S. Tomisato, K. Chiba, and K. Murota, “Phase error free LINC modulator,” Electron. Lett., vol. 25, no. 9, pp. 576–577, Apr. 1989.
[5] L. Sundstrom and M. Johansson, “Effect of modulation scheme on
LINC transmitter power efficiency,” Electron. Lett., vol. 30, no. 20,
pp. 1643–1645, Sept. 1994.
[6] F. H. Raab, “Efficiency of outphasing RF power-amplifier systems,”
IEEE Trans. Commun., vol. COM-33, pp. 1094–1099, Oct. 1985.
[7] T. Hornak and W. McFarland, “Vectorial signal combiner for generating
an amplitude modulated carrier by adding two phase modulated constant
envelope carriers,” U.S. Patent 5 345 189, Sept. 1994.
[8] T. Gabara and W. Fischer, “An integrated system consisting of an
818 adiabatic-pps multiplier powered by a tank circuit,” in IEEE Int.
Solid-State Circuits Conf. Tech. Dig., vol. 387, New York, 1995, pp.
316–317.
[9] G. Hanington, P. Chen, and P. Asbeck, “A 10-MHz DC-DC converter for high-efficiency power amplification in RF transmitters,” in
UCSD/IEEE Conf. Wireless Commun., 1998, pp. 74–78.
Fig. 6. Measured variation of circuit efficiency as a function of input power
level for 50% AM at 10 kHz and CDMA at 1.25 MHz at a carrier frequency
of 1.96 GHz, for Vsup = 3, 4, and 5 V.
by a dc–dc converter. One of the main goals in this design was
to optimize the power re-use efficiency at the 1.96-GHz band,
and to investigate the effect of variations in the power-supply
voltage on efficiency. However, from a practical standpoint, a
dc–dc converter would be desirable to provide the proper supply voltage to the power amplifiers in any given system. This
approach has recently been developed for the case of dynamic
power-supply variations in linear power amplifiers [9].
One issue that has not been addressed to this point is the inherent problem in the LINC architecture due to amplitude and
phase imbalance between the two power amplifiers. Both of
these imbalances will result in a lower amplitude modulation
index signal on the output to the antenna and the detector. A
small degradation in re-use efficiency will be noticed in the
system due to this effect.
IV. CONCLUSION
A new technique has been presented for the power-added
efficiency improvement of outphased microwave power amplifiers, which involves the “recycling” of a portion of the wasted
power in the summing hybrid back to the power supply. The
circuit achieved a peak power re-use efficiency of over 60%,
and the efficiency was greater than 40% over more than a
decade of input power at 1.96 GHz. This technique promises to
improve the overall efficiency of outphased microwave power
amplifiers, making them more attractive for future generations
of highly linear amplifiers.
ACKNOWLEDGMENT
The authors acknowledge several valuable discussions
with Prof. D. Rutledge, California Institute of Technology,
Pasadena, Dr. M. Yoder, Office of Naval Research, Washington, DC, G. Hannington, University of California at San
Diego, La Jolla, and K. Gard, Qualcomm Inc., San Diego, CA.
Robert Langridge (S’83–M’86) was born in Cedar
Rapids, IA, on December 4, 1963. He received
the B.S. degree in electrical engineering from Iowa
State University, Ames, in 1986, and is currently
working toward the M.S. degree in electrical engineering at the University of California at San Diego,
La Jolla.
From 1995 to 1997, he was with Glenayre Electronics Inc. Quincy, IL, where he was involved
in the RF front-end design of high-dynamic-range
low-noise receivers for 900-MHz two-way paging
systems. From 1993 to 1994, he was with Sciteq Electronics Inc., San
Diego, CA, where he was involved with phase-locked loop (PLL) frequency
synthesizers for various applications, which included the fractional-N techniques. From 1987 to 1993, he was with IFR Systems Inc., Wichita, KS,
where he was involved with several RF and analog design projects related
to communication service monitors and general-purpose spectrum analyzers,
including an RF signal generator, various frequency synthesizer projects, and
spectrum-analyzer IF design. From 1986 to 1987, he was with Boeing Military
Airplane Company, Wichita, KS, where he was involved with digital and RF
circuit design and analysis. He holds one patent.
Mr. Langridge is a member of Eta Kappa Nu.
Todd Thornton received the M.S.E.E. degree from the Georgia Institute of
Technology, Atlanta, the B.S.E.E. degree from Drexel University, Philadelphia, PA, and is currently working toward the Ph.D. degree at the University
of California at San Diego, La Jolla, where his research is concentrated on
amplifier linearization techniques.
He is currently a Senior RF Engineer at REMEC Wireless, San Diego,
CA, where he designs millimeter-wave and microwave transceivers, lownoise amplifiers, high-power RF amplifiers, and linearization circuitry for
RF amplifiers. He was a Staff Engineer at ANTEC, Atlanta, GA where he
designed fiber-optic transmitters and receivers. He was also a Project Engineer
at Electromagnetic Sciences, Atlanta, GA where he designed solid-state power
amplifiers for satellite communications. He has also worked as a Junior Level
Designer on RF amplifiers at SGS Thomson Microelectronics, Microcom Inc.,
GE Astro Space, and American Electronic Laboratories. He is a veteran of the
U.S. Marine Corps, where he was a Electronic Countermeasures Specialist.
Peter M. Asbeck (M’75–SM’97), for photograph and biography, see this
issue, p. 1437.
REFERENCES
[1] H. Chireix, “High power outphasing modulation,” Proc. IRE, vol. 23,
pp. 1370–1392, Nov. 1935.
Lawrence E. Larson (S’82–M’82–SM’90), for photograph and biography,
see this issue, p. 1403.